Fetching the paper…

AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration · Around